We use fmtlib (fmt::fmt cmake target) in our software, by downloading and building it with the rest of our software (cmake FetchContent and friends...).
We encountered a build failure using the newest humble release, which i traced back to a rouge fmt::fmt target appearing in our build (preventing us from building the latest fmtlib version ourself), transitively through rclcpp, rcl, rcl_logging and spdlog.
I'm not sure which component should or should not propagate that dependency upwards, but i think !90 is the particular change causing our problem here. I also found !60, which seems to do the exact opposite and seems convincing as well :smile:
Feel free to close this as not an issue with rcl_logging, but i wanted to open an issue here as this might be a regression for more users than just us...
We use fmtlib (
fmt::fmtcmake target) in our software, by downloading and building it with the rest of our software (cmake FetchContent and friends...).We encountered a build failure using the newest humble release, which i traced back to a rouge fmt::fmt target appearing in our build (preventing us from building the latest fmtlib version ourself), transitively through rclcpp, rcl, rcl_logging and spdlog.
I'm not sure which component should or should not propagate that dependency upwards, but i think !90 is the particular change causing our problem here. I also found !60, which seems to do the exact opposite and seems convincing as well :smile:
Feel free to close this as not an issue with rcl_logging, but i wanted to open an issue here as this might be a regression for more users than just us...